mindmap: Spatial Memory in Deep Feature Maps for 3D Action Policies

A Spatial Retrieval-Augmented Generation system for latent world models, designed for embodied spatial intelligence in robotics, autonomous navigation, and embodied AI. Features ROS2 integration, real-time inference @ 25Hz, and complete robot build guide.

Spatial field location memory for document processing pipelines. Learns field positions, validates extractions, identifies document types by layout, detects drift, and monitors template health. Zero dependencies, pure Python.

A Machine Learning Aware Spatial Data Re-partitioning Framework for Spatial Datasets

Spatial Metaphors for LLM Memory: A Critical Analysis of the MemPalace Architecture — Full paper, benchmarks, experiment data, and reproduction scripts

Control architecture for an agent that perceives, plans, learns, and acts in a 3D MMORPG. Evolved from a monolith through reactive rules, state machines, and utility scoring to GOAP planning. JPS/A* pathfinding, DDA line-of-sight, Bayesian adaptation. Pure Python 3.14 (free-threaded), zero dependencies.

mindmap

Spatial memory layer for AI agents — RGB-D / point clouds to a persistent, queryable, LLM-native 3D scene graph. Mem0 for 3D space.

Spatial scene graphs and belief state for embodied agents